Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Real-Time

By Chantat Eksombatchai, Pranav Jindal, Zitao Liu, Rahul Sharma, Charles Sugnet, Mark Ulrich, Jure Leskovec. 2017

229 citations recorded by the source corpus.

User experience in modern content discovery applications critically depends on high-quality personalized recommendations. However, building systems that provide such recommendations presents a major challenge due to a massive pool of items, a large number of users, and requirements for recommendations to be responsive to user actions and generated on demand in real-time. Here we present Pixie, a scalable graph-based real-time recommender system that we developed and deployed at Pinterest. Given a set of user-specific pins as a query, Pixie selects in real-time from billions of possible pins those that are most related to the query. To generate recommendations, we develop Pixie Random Walk algorithm that utilizes the Pinterest object graph of 3 billion nodes and 17 billion edges. Experiments show that recommendations provided by Pixie lead up to 50% higher user engagement when compared to the previous Hadoop-based production system. Furthermore, we develop a graph pruning strategy at that leads to an additional 58% improvement in recommendations. Last, we discuss system aspects of Pixie, where a single server executes 1,200 recommendation requests per second with 60 millisecond latency. Today, systems backed by Pixie contribute to more than 80% of all user engagement on Pinterest.

Subfields: Computer Networks and Communications, Computer Vision and Pattern Recognition, Information Systems, Recommender Systems, ML Systems & Hardware

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

1. Yong Li
2. Xiangnan He
3. Tat-seng Chua

In Computer Vision and Pattern Recognition

1. Yong Li
2. Chen Gao
3. Depeng Jin
4. Fei Sun
5. Bin Cui
6. Shiwen Wu
7. Fei Sun
8. Bowen Jin

In Information Systems

1. Yong Li
2. Chen Gao
3. Xiangnan He
4. Depeng Jin
5. Fei Sun
6. Bin Cui
7. Shiwen Wu
8. Fei Sun

In Recommender Systems

1. Yong Li
2. Chen Gao
3. Fei Sun
4. Bin Cui
5. Shiwen Wu
6. Fei Sun
7. Depeng Jin
8. Xiangnan He

In ML Systems & Hardware

1. Bin Cui
2. Xiangnan He

Citations in the corpus

• Multi-behavior Recommendation with Graph Convolutional Networks (2020, methodology): “Some works [1, 8, 12, 25] apply random-walk based methods to graph-based recommendation.”
• Graph Neural Networks in Recommender Systems: A Survey (2020, background): “In industry, GNN has also been deployed in web-scale recommender systems to produce high-quality recommendation results [32, 113, 189].”